Arnott, D., & Pervan, G. (2005). A critical analysis of decision support systems research. Journal of Information Technology, 20, 67‐87.

More summarizing, another Arnott and Pervan work, this one from 2005. These guys spent a lot of time reading the DSS literature.

Arnott and Pervan (2005) appear to get one thing… well, one should not say wrong in these genteel academic confines, so let’s just say there’s an incomplete hypothesis. They point to the predominace of theory building and theory testing and the lack of theory refinement. Throughout the sample period, 90% of published DSS research has dealt with theory building and testing. Testing is up and building down, but still refinement remains a rarity. Four decades, and we’re not refining theory yet? What gives?

Arnott and Pervan postulate that “new DSS movements, especially EIS, data warehousing, and business intelligence” may have given many researchers cause to return to square one (78); however, this observation appears to offer only a partially sufficient explanation of this phenomenon. Arnott and Pervan find that data warehousing is the only one of those three areas that has experienced an increase in relevant article publication, and those three areas together still constitute less than a quarter of the total sample (Table 7, p. 77). A quarter of the articles skewing the sample toward theory building and testing seem unlikely to provide sufficient weight to limit theory refinement to just 3.3% of the sample. Plus, Pervan and Arnott identify the very absence of EIS and data warehousing research as a key shortcoming researchers must redress to increase the relevance of the field.

Another strong indicator of the lack of relevance in DSS research is the remarkable absence of any clear definition of primary clients and users in published articles. This lack of context fits with the apparent predominance of the positivist paradigm in DSS research: one could hardly conduct interpretivist or design research without interacting with the clients and/or users. Conceivably one might defend the absence of clearly defined clients and users from a given article as an effort to generalize findings beyond a specific context; however, practitioners want a clear picture of what a decision support system could do for their organizations. Rather than leaving them to reconstruct the client–system–user dynamic that might have produced the research results in the first place, researchers would do better to enrich their studies with those very details so practitioner readers may immediately access the relevance of that research to specific applications.

Arnott and Pervan go so far as to cast DSS research as experiencing a “crisis” of relevance (p. 82). In their subsequent 2008 paper, while they still focus on the same dataset, the authors appear to see no clear abatement of that crisis. The recommendations they make in this 2005 article—breaking free of research inertia to focus less on PDSS/GDSS and more on DW/BI; conducting more case studies; strengthening design research; and moving beyond the theoretical basis from the behavioral decision theory of Herbert Simon to forge a stronger, more interdisciplinary framework incorporating “contemporary research in psychology, management, and related fields” (p. 83)—appear to remain worthy of consideration today.

Advertisements